It is becoming more and more normal to have separately actuating compo-nents, see Bjerknes [2012], rather than large linkage systems. The reason for these are numerous but one of the main reasons is the increased complexity that rules out larger linkage systems. Also mathematical model are easier to follow with individually actuated joints, see section 7 on page 63, and the flexibility of such system meets today’s high standards.
Since our robot is electric we found out that hydraulically actuated joints would only introduce delay, increase complexity and reduce efficiency, and the solution is to have electrically actuated joints. Normally such rotating joints are served by a servo motor and this is solution that the author has seen the benefits of live, see Bjerknes [2012].
Figure 3.5: An exploded view of the steering components [Blomberg, 2014].
3.5.1 Electric Servomotors
”A Servo Motor is defined as an automatic device that uses an error-correction routine to correct its motion. The term servo can be applied to systems other than a Servo Motor; systems that use a feedback mechanism such as an encoder or other feedback device
CHAPTER 3. STEERING SYSTEM 29
to control the motion parameters. Typically when the term servo is used it applies to a ’Servo Motor’ but is also used as a general control term, meaning that a feedback loop is used to position an item.” [Anaheimautomation, 2013]
They are often found in industrial applications were precise operations are needed and are generally more expensive than a stepper motor that can be used for simpler, yet similar tasks [Anaheimautomation, 2013].
There are two main types of electric motors on the market today, and the cheapest and easiest to control are brushed motors that has shorter lifespan and poorer efficiency than brush-less motors, that delivers more torque at low speeds and requires more sophisticated motor controllers. In our robot design we chose brush-less motors to save weight and have good low rpm torque.
An expensive technology becoming more and more popular is integrated servomotors, and the idea here is to add simplicity in the physical set-up of the motor and motor controller by integrating the motor controller inside the motor making them one unit. This reduces volume and offers better protection to the servo system as fewer connections are needed. The control of such a system can be described more like plug and play compared to separate components, as described in section 5.1.4 on page 46.
Integrated servomotor
The motor that were chosen is a Danish high end brush-less integrated JVL MAC 141 servo motor of 134 Watts, capable of delivering 0.48 nm nomi-nal and 1.59 nm peak torque [JVL Datasheet, 2014]. It is fitted with an CANopen communication protocol interface, MAC00-FC4, with m12 indus-trial connectors to receive and send messages over the CAN bus-cable, via the PEAK CAN-USB adapter, connected to the laptop that runs the Robot Operating SystemROS Hydro Medusa.
The mode that servo motors is going to operate in is called position mode and more information on this mode is found in JVL Manual [2014]. In this mode the servo motors follows the commended positions from ROS, which is running the Ackerman equations found in section 7.6 on page 75, and if a servomotor senses that it isn’t moving according to the commanded position, it will apply force to get in the correct position.
3.5.2 Reduction Gears
The steering torque needed to operate the steering system varies much, and the best solution would have been to do field tests, that identified the needed
30 CHAPTER 3. STEERING SYSTEM
(a) The Integrated JVL MAC 141 134 watts servo motor chosen for our mobile agricultural robot.
(b) The APEX DYNAMICS AB060 two stage planetary gear with 60:1 ratio chosen for the servo motors.
Figure 3.6: Some of the mechanical components used in the steering system.
torque to turn the wheels of the robot under different conditions, but the climate in Norway at this stage didn’t allow for this. Instead a similar method to Madsen and Jakobsen [2001] is used, were one assumes that the wheel is stationary on dry concrete, and then uses a simple formula to calculate the needed torque to turn the wheel, see Grimstad [2014] for more details.
Obliviously the lowest steering torque demand is present when the robot is moving on hard surface, and the highest steering torque demand is when the robot is stand still situated in deep mud. Since our robot is likely to operate in both conditions, and by including the fact that efficiency goes down if gear ratio is increased, we must find a gear ratio and gear efficiency that gives us fast enough steering rate, while at the same time offering high enough torque. Since the servo motor is capable of delivering three times the nominal torque under a short period, see section 3.5.1 on the preceding page, the muddy conditions will be covered by this feature.
CHAPTER 3. STEERING SYSTEM 31
Worm Gears
Worm gears are cheap and readily available, and they can be be delivered in a wide range of gear ratio. They can be advantageous if a 90 degree bend on the gear is needed and there is limited space. The downsides of these gears are their efficiency, and it is not uncommon that it is under 50%. This is undesirable on the robot if the steering system is activated often, as precious battery energy is transferred to heat via friction. Another feature of gears in general is that they are self locking if efficiency is 50 % or under SEW-EURODRIVE [2014], this means that the gear can hold the same torque that it can provide. Self locking worm gears are used in the Mobile Robot for Weeding shown infigure 2.5 on page 13 and Madsen and Jakobsen [2001]
reported problems with this feature.
Planetary Gear
Planetary gear, also called Epicyclic gear, are regarded as one of the most efficient gear types according to [SEW-EURODRIVE, 2014]. They can be delivered in one stage or multi stage, where the later is several planetary gears stacked together to increase gear ratio. They are also significant lighter than a similar worm gear.
A APEX DYNAMICS AB060 two stage planetary gear with 60:1 ratio is chosen for the NMBU mobile robot. This gear has efficiency of >94%
according to Grimstad [2014], and will mounted directly above steering axle under servo motor. This gives us nominal steering torque of 0.48nm×60× 0.94 = 27nm and a peak steering torque of 1.59nm×60×0.94 = 90nm.
Remarks
The steering components presented in section 3.5 on page 28, form a steering actuator system which we believe isstate of the art at the moment, and the reason for this is that the components gives us a low volume, small footprint, high efficiency, high precision, high torque, coaxial steering actuator.